I-Projection Results

Steven

2025-11-19

Q1.) Do projections move closer to the positive end of the axis over time?

Note. Data are aggregated at the assessment level and coefficients are unstandardized.First person sing. = means of I, me, and my projections for each participant.

first_person_sing_valence_combined

first_person_sing_valence_combined individual variation

first_person_sing_valence_combined spline fit

First-person sing. able projections (am, feel, & combined)

Note. Data are aggregated at the assessment level and coefficients are unstandardized.First person sing. = means of I, me, and my projections for each participant.

first_person_sing_ability_combined

first_person_sing_ability_combined individual variation

first_person_sing_ability_combined spline fit

Q2.) Do our projection measures predict changes in internalizing symptoms over time?

Results are condensed to be for the first-person singular valence and ability projections only (combined).

Ability Projection

Non-disaggregated
Disaggregated

Q3.) Are these predictions robust to other linguisitic features?

Results are condensed to be for the first-person singular valence and ability projections only (combined).

Valence Projections

Non-disaggregated (Valence)
Disaggregated (valence)

Ability Projections

Non-disaggregated (Ability)
Disaggregated (ability)

Q4.) Can we cluster based on projection scores?

Good-bad

Map Change over time in each cluster

Cluster 1
Cluster 2
Cluster 3
Cluster 4
Plot trajectories for each cluster

# Standardization method: pseudo

Parameter          | Std. Coef. |       95% CI
----------------------------------------------
(Intercept)        |       0.00 | [0.00, 0.00]
daysSinceFirstText |       0.06 | [0.05, 0.06]

# Standardization method: pseudo

Parameter          | Std. Coef. |       95% CI
----------------------------------------------
(Intercept)        |       0.00 | [0.00, 0.00]
daysSinceFirstText |       0.07 | [0.06, 0.08]

# Standardization method: pseudo

Parameter          | Std. Coef. |        95% CI
-----------------------------------------------
(Intercept)        |       0.00 | [ 0.00, 0.00]
daysSinceFirstText |  -3.67e-03 | [-0.01, 0.00]

# Standardization method: pseudo

Parameter          | Std. Coef. |         95% CI
------------------------------------------------
(Intercept)        |       0.00 | [ 0.00,  0.00]
daysSinceFirstText |      -0.02 | [-0.03, -0.02]

Differences in Symptoms Across Clusters

Baseline Symptoms
Final Symptoms
Change in Symptoms

Main Effects
Pairwise Comparisons

Main Effects
Pairwise Comparisons

Controlling for baseline symptoms

Main Effects
Pairwise Comparisons

Plot

Do clusters moderate?

model = lmer(Internalizing ~ first_person_sing_valence_combined * clusterComb + daysSinceFirstText + (1|room_id), data = data_cluster)

Table

Plot

All plots together

Able-Unable

Map Change over time in each cluster

Cluster 1
Cluster 2
Cluster 3
Cluster 4
Plot trajectories for each cluster

# Standardization method: pseudo

Parameter          | Std. Coef. |       95% CI
----------------------------------------------
(Intercept)        |       0.00 | [0.00, 0.00]
daysSinceFirstText |       0.10 | [0.08, 0.11]

# Standardization method: pseudo

Parameter          | Std. Coef. |       95% CI
----------------------------------------------
(Intercept)        |       0.00 | [0.00, 0.00]
daysSinceFirstText |       0.06 | [0.06, 0.07]

# Standardization method: pseudo

Parameter          | Std. Coef. |       95% CI
----------------------------------------------
(Intercept)        |       0.00 | [0.00, 0.00]
daysSinceFirstText |       0.06 | [0.05, 0.06]

# Standardization method: pseudo

Parameter          | Std. Coef. |         95% CI
------------------------------------------------
(Intercept)        |       0.00 | [ 0.00,  0.00]
daysSinceFirstText |       0.07 | [ 0.06,  0.08]